[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Apache Tika for #499606



Hi Thomas,

Initially, i was trying to use git-buildpackage, then saw they are using quilt; so i swapped to use topgit, just to try out. Yes, i do agree and feel separate to different branch a bit strange. Because when i clone back the package to other machine. I could not see the debian/0.7 branch at all, git branch don't show it. (May be i don't know how to use git ...) Would like to try your tnt if available.

To continue ... i think fixing the dependencies is the priority. The other maven repos missed package, i think better put up a bug report first, because at this stage i have no knowledge on ant and maven ... i.e. build.xml and pom.xml

Questions are
1. What is the debian Java Package direction; are all packages should built with Maven finally?
2. For build.xml and pom.xml ... i need some tutorial how to write or generate ... best is somebody help to write for tika so that i can learn from it.

Thanks.
~Pat.

On Wed, Jun 23, 2010 at 10:38 PM, Thomas Koch <thomas@koch.ro> wrote:
Hi Patrick,

nice to hear from you again! :-) Please subscribe also to the debian-java list
as described here:
http://wiki.debian.org/Teams/JavaPackaging
The pkg-java-maintainers list is not for discussions but for some kind of
status mails.

More replies inline:

Patrick Yuen:
> Dear All,
>
> After reading http://wiki.debian.org/HowToPackageForDebian directed by
> Thomas Koch. I have drafted apache tika in git format at
> http://git.debian.org/?p=collab-maint/tika.git;a=summary
>
> It was done after referenced solr and ivy. Up to this point, i do need some
> help or ideas how to continue.
>
> Please download and have a quick look.
> git clone http://git.debian.org/git/collab-maint/tika.git
> git checkout debian/0.7

I'd strongly advise against using topgit:
http://www.koch.ro/blog/index.php?/archives/139-tnt-is-not-topgit.html

'master' should be the branch containing the debian/ dir. If you're using git-
buildpackage, then this is also the default.

> 1. Missing Dependencies is written in README.Debian ... since i could not
> find from debian packages repository, i guess it might be missing or may be
> it is there in different name ?
>
> Full dependency is listed at
> http://tika.apache.org/0.7/gettingstarted.html. Three missing dependencies
> are ...
>
> > poi-ooxml-3.6.jar
> > poi-ooxml-schemas-3.6.jar
> > geronimo-stax-api_1.0_spec-1.0.jar
Torsten Werner, a DD, provides an autogenerated list of all Java Classes in
all JARs in all Debian Java packages:
http://ftp-master.debian.org/~twerner/jar-content.txt.bz2

(Would be fine to have this as API somedays...)

I checked the file and indeed it seems no package in Debian contains these jar
files. I just created this wiki page for you:
http://wiki.debian.org/JavaPackagingFAQ

> 2. MavenBuilder - http://wiki.debian.org/Java/MavenBuilder ... Does it
> suppose to get all the dependencies from /usr/share/maven-repo or i need to
> do something, because some packages are in debian but not in maven-repo ...
> refer README.Debian
AFAIK MavenBuilder can only use those dependencies that are available in
/usr/share/maven-repo. There are many Debian-Java packages which are not
availabe there. You're very welcome to help there too. :-)

> 3. Need help on build.xml and pom.xml ... not familiar on ant and maven at
> all.
Could you please be more specific?

> 4. Is the package name correct ? tika or libtika-java ? it does have a gui
> application after maven built.
You should build at least three binary packages: libtika-java to be reused
e.g. by SOLR, tika-gui (or another name) and libtika-java-doc for the javadoc
and any other documentation that may be included in tika.

> Thanks.
You too.
>
> ~Pat.

Thomas Koch, http://www.koch.ro


Reply to: